Look at this Google-cached (pulled down) PlayStation 3 page

[DeadmeatGA]

...

I really don't care about hardware specs anymore, there is no such thing as an "ultimate" hardware and somebody else is going to beat you 6 month later no matter what you do. If you think PSX3 will go unchallenged for years, then you are sadly mistaken. Which ever comes out 6 months after PSX3 will beat it, it is a fact.

What I really care about nowadays is the programmability of such massively parallel architecture; one theoritical teraflop is nice, but can an average programmer code for it??? The answer is no. Programmability comes first, performace potential is the secondary consideration.

 

[notAFanB]

What I really care about nowadays is the programmability of such massively parallel architecture; one theoritical teraflop is nice, but can an average programmer code for it??? The answer is no. Programmability comes first, performace potential is the secondary consideration.

so in your opinion CELL has no chance of scaling it's theoreticle peak performence and SONY execs have been conned my a madman with delusiosn of granduer.

or something to that effect.

 

[Fafalada]

Hello Deadmeat 8)

 

[DeadmeatGA]

...

so in your opinion CELL has no chance of scaling it's theoreticle peak performence

There are people who could do it, experienced MPI programmers. But your average game programmer does not fit the job description.

SONY execs have been conned my a madman with delusiosn of granduer.

Kutaragi was powerful enough to kill the handheld developed by Sony headquarter back in 1998, simply because it was developed without his approval. Kutaragi is the current top dog in power struggle within Sony group.

 

[DeadmeatGA]

...

Hello Deadmeat

Hi, you should be the one to bitch most about the PSX3 architecture here. Afterall, it is you who will be struggling on this beast, not us.

 

[notAFanB]

There are people who could do it, experienced MPI programmers. But your average game programmer does not fit the job description.

alright then one last question. if there was an API developed for CELL whom would likely be working on it? and chances of success? are the above MPI coders any help at all.

 

[DeadmeatGA]

...

if there was an API developed for CELL whom would likely be working on it? and chances of success? are the above MPI coders any help at all.

MPI is merely a socket. It does not make software automatically parallel, parallelism has to be hand-coded in between processes using sockets(MPI).

And I don't consider games to be the kind of software that rends itself to parallelism, there are too much inter-dependencies between game data structure to go MPP. Non-interactive graphics rendering, yes. Games, no.

 

[Panajev2001a]

Kutaragi is the current top dog

For Sony as a whole, it is a good thing he is the top dog IMHO

 

[Panajev2001a]

Re: ...

heh, the saturn was a rush job I think so that comparison is a little unfair.

The problem is that Japanese electronic industry is very hardware centric and software guys have little say in how hardware is designed. This is why we have flawed designed like PSX2 and PSX3, all shooting for big marketting hype numbers at the expense of programmability. If PSX2 failed then PSX3 would have been a nicer machine(Was the case with DC and GC, both developer friendly machines), But since developers did not protest PSX2's "unusual" architecture, Kutaragi struck back with an even worse design, the PSX3. This is why SOFTWARE GUYS and not hardware guys should be in charge of hardware development.

BEEEP!

It is never software guys OR hardware guys designing the Hardware... you have to go with both groups in synchrony that is the difference between a good processor and a so-so processor...

If you let the software people design the Hardware you would get enormous delays as the circuit engineers and the rest of the EE guys are pulling their hair out trying to tell those num-nats upstairs that their design is unfeasible...

Hardware guys and Software guys work together in succesful designs ( one of the strengths of the former Alpha )...

You are weird Deadmeat, in the sense that one minute Sony begged for Cell and the next minute Kutaragi and his team of pure EE-we-do-not-know-anything-about-software ( as you call them ) people designed Cell on their own and that is why it will be a impossible to code for...

Cell has its roots in IBM's R&D labs and all three companies sent legions of software engineers in the project... the guys who defined Cell's architecture and its ISA ( we have a group of "magic five" from IBM plus top dogs at Sony and Toshiba ) are not incompetent engineers like you make them sound... they were worldwide recognized senior MPU architects.

so parrellism works is MS is working on it? is it unfeasable to write DX stlye API's of other parrellell architectures

It is feasible, but Sony doesn't have the experience to do it.

You mean Sony+IBM+Toshiba ( nevermind Sony has been very active following DirectX and solutions like Stanford Shading Language ) do not have enough experience ?

is this severly limited to the way CELL appears to be heading?

CELL was not developed with programmability in mind; the guy who masterminded CELL is an electrical engineer, not a programmer. He simple doesn't know.[/quote]

See above...


It is funny you mention shooting for high performance without thinking about ease of use and link the Saturn to this approach and then you show how it fits in the PlayStation 2 and PlayStation 3 line of development...

I do not know, but the Saturn shows me a faled attempt at pushing for fast performance... nobody with a sane mind would have thought about a multiprocessing set-up like that, it was an after-thought trying to gain last minute performance and its design shows it... the Hardware itself is a barrier to synchronize execution resources... the SCU's DSP steps over memory transaction from either SH-2 to main RAM as well as any inter-processor communication ( the SH-2s have little cache and need to use main RAM to pass data back and forth )... not a very neat architecture for 3D IMHO.

 

[DeadmeatGA]

....

If you let the software people design the Hardware

What I meant is that software guys design the "architecture", hardware guys come up with the "implementation" of the architecture...

one minute Sony begged for Cell

Sony didn't beg for CELL, I don't think Sony headquarter even likes CELL. It's all Kutaragi's doing..

and the next minute Kutaragi and his team of pure EE-we-do-not-know-anything-about-software ( as you call them ) people designed Cell on their own and that is why it will be a impossible to code for...

Here is my version of CELL chronology.

1. While contemplating about next PSX in his head, Kutaragi visits IBM on a business trip and sees Blue Gene CELL program in 1999.
2. Kutaragi is impressed with IBM"s work and decides that this is the direction his next playstation should be going.
3. Instead of adapting pure CELL, Kutaragi negociates to CELLtize PSX2 VUs to claim big performance numbers.
4. Kutaragi and IBM announces the VU CELL on March 2001 and the rest is history.

Now, we have the worst combination of two previous architectures without the abstraction to remedy the situation; the explicit parallelism of BlueGene CELL with assembly coding of PSX2 VU. Developers are going to have one hell of a bumpy ride with their PSX3 development.

I do not know, but the Saturn shows me a faled attempt at pushing for fast performance...

But the Saturn architecture did influence PSX2 greatly.

nobody with a sane mind would have thought about a multiprocessing set-up like that

Both are designed on same architectural principle, a distributed processing architecture built around a powerful DMA engine shifting data around. The thinking behind PSX2 architecture is very different from PSX1 and shows a strong Saturn influence.

, it was an after-thought trying to gain last minute performance and its design shows it...

I can clearly tell one of PSX2's VUs was a last minute add on in an attempt to boost geometric processing numbers against DC. I suspect it is VU1.

 

[Vince]

Re: ....

Deadmeat, we need to get past these psycologically induced illusions that plague your conscious thinking. I suggest you start here: http://www.nsf.gov/od/lpa/news/publicat/frontier/6-96/6illusio.htm

What I meant is that software guys design the "architecture", hardware guys come up with the "implementation" of the architecture...

What? No they don't - hardware companies have specific teams that do nothing but advanced product placement. When was the last time a software engineer told IBM or Intel how it's going to be done?

Sony didn't beg for CELL, I don't think Sony headquarter even likes CELL. It's all Kutaragi's doing..

See above link to psychological diagnosis.

EDIT: You do realize that Cell, if it suceeds, will be a turning point in the companies history. Have you even heard Kunitake Ando talk in the last year?

Here is my version of CELL chronology.

This will be fun.

. While contemplating about next PSX in his head, Kutaragi visits IBM on a business trip and sees Blue Gene CELL program in 1999.
2. Kutaragi is impressed with IBM"s work and decides that this is the direction his next playstation should be going

Kutaragi went to IBM first in 2000, having proposed his idea for a radically new architecture based on distibuted computing and data sharing to IBM after negotiations with Toshiba. Blue Gene hardly had the infleunce you think - I (among others) initially anticipated there was a link due to the name, goal, et al. What is subsequently found is that the Cellular Computing theory is hardly embodied in the praxis that is PS3's Cell. It's neither small, nor cheap (in logic). Just compare them, Cell is much more coarsly grained in execution units and hierarchial - hardly Blue Gene like.

3. Instead of adapting pure CELL, Kutaragi negociates to CELLtize PSX2 VUs to claim big performance numbers.

There is no architectural commonality between then threw phsycial IP resuse AFAIK. This is all your biased head filling in all you don't know (which is alot my friend).

. Kutaragi and IBM announces the VU CELL on March 2001 and the rest is history.

Umm, this is true - but the only part of your timeline that you didn't think up... go figure.

Now, we have the worst combination of two previous architectures without the abstraction to remedy the situation; the explicit parallelism of BlueGene CELL with assembly coding of PSX2 VU. Developers are going to have one hell of a bumpy ride with their PSX3 development.

What? Making blanket statements with no grounds in reality are reserved for Chap on these boards...

But the Saturn architecture did influence PSX2 greatly.

< Hits head on wall > Ketchup is red, Salsa is red... Ketchup influenced Salsa. Good times....

Both are designed on same architectural principle, a distributed processing architecture built around a powerful DMA engine shifting data around. The thinking behind PSX2 architecture is very different from PSX1 and shows a strong Saturn influence.

It's different? the Graphic Synthesizer is built on the reuse of PS1's IP.

The Emotion Engine hardly shows a "strong Saturn Influence." Try the work of Diefendorff, or Mike Raam, or even go back the the initial IEEE paper by Kutaragi (A Microprocessor with 128b CPU, 10 Floating-Point MACs, 4 Floating-Point Dividers, and MPEG2 Decoder) and tell me where it accredits anything. Similarly look to papers by SCE and Toshina engineers such as Sukuoki, Oka, Ide, Tago and Murakami.

I fail to see it... want to know why? Because your full of shit.

I can clearly tell one of PSX2's VUs was a last minute add on in an attempt to boost geometric processing numbers against DC. I suspect it is VU1.

I can clearly call Bullshit.

K. Kutaragi, et. al., "A Microprocessor with 128b CPU, 10 Floating-Point MACs, 4 Floating-Point Dividers, and MPEG2 Decoder", in 1999 IEEE International Solid-State Circuits Conference, Feb 15-17 1999, San Francisco CA

A. Kunimatsu et al., “5.5 GFLOPS Vector Units for Emotion Synthesis,â€￾ Hot Chips 11 Conf. Record, Aug. 1999, pp. 71-82.

N. Ide et al., “2.44 GFLOPS 300MHz Floating-Point vector Processing Unit for High-Performance 3D Graphics Computing,â€￾ Proc. European Solid-State Circuits Conf. of CMOS logic VLSI chips. (ESSCIRC 99), Editions Frontieres, Gif-sur-Yvette, France, ISBN 2-86332-246-X, 1999.
pp. 106-109.

You do realize the entire Emotion Engine was presented in complete in early-1999. With a development cycle of a conservative one year... that's late-1997/1998. You're worse than Chap, atleast he's funny.

 

[Katsura]

Yo Deadmeat,

Where is my doomed GBA (as a collector's item for such a short lived product, of course), my PSX beating 3D mobile phones, and oh, how is Super CRYO?

Greetings.

 

[London Geezer]

one word:

 

[DeadmeatGA]

...

Deadmeat, we need to get past these psycologically induced illusions that plague your conscious thinking. I suggest you start here:

Do you not realize that above line constitutes a personal attack and you are now subject to an immediate ban and full profile revelation?? I will be misssing you....

EDIT: You do realize that Cell, if it suceeds, will be a turning point in the companies history. Have you even heard Kunitake Ando talk in the last year?

Nope. CELL won't change IBM, CELL won't change Sony, CELL will only affect SCEI. Come on, CELL powered TV sets and DVD players??? Who else is going to put SCEI's proprietary CELL in their product? Matsushita? JVC? Samsung? For what purpose? Will Sony want to carry a substantial cost burden that will serve little purpose??? Will consumers go gaga over $2000 TV sets that will supposely accelerate their PSX3 games by 10%??? Will IBM dump its X86 and POWER servers in favor of CELL servers??? Use some common sense here.

CELL, if successful, will make somewhat pretty game console. It won't change the world. Let's wake up from the dream and smell the coffee here.

Kutaragi went to IBM first in 2000

Thanks for the correction. That's when Kutaragi saw Blue Gene for the first time.

having proposed his idea for a radically new architecture based on distibuted computing and data sharing to IBM

That's odd, since IBM came up with the term Cellular Computing in 1999 and Blue Gene project began right after Deep Blue. If Kutaragi's CELL was an independent idea why go to IBM and pay big bucks to have it developed??? After all, CELL is an IBM baby and SCEI is a paying customer, much like EE was a Toshiba baby and SCEI was the paying customer.

Blue Gene hardly had the infleunce you think - I (among others) initially anticipated there was a link due to the name, goal, et al.

I already see the connection there. Big Blue doing research on something called "Cellular Computing", and Kutaragi san wanting a piece of it....

Just compare them, Cell is much more coarsly grained in execution units and hierarchial - hardly Blue Gene like.

This is why I said that Kutaragi wanted to CELLtize EE VUs after seeing Blue Gene.

You do realize the entire Emotion Engine was presented in complete in early-1999. With a development cycle of a conservative one year... that's late-1997/1998. You're worse than Chap, atleast he's funny

Dreamcast was completed by early 1998 and Kutaragi had its full spec by then. Of course he wanted to beat Sega numbers by a large margin, and this is why he ordered Toshiba to throw in VU1.

 

[chaphack]

OT..or not! BUT after what Xbox has given us, i expect no less than full HDTV support for next gen consoles.

60fps for 480i, 60fps for 480p, 60fps for 720p, 60fps for 1080i, 60fps for 1080p(!!) Say bye bye to 32bpp too!

I think we should drop the 640 32bpp calculations. Sure they might not be neccessary on average, but we should get more options. It is next gen afterall.

I am sure if Xbox follows the PC route again, it will support them all and more. Heres hoping Sony dont butcher the PS3 options too much with all those high cost investments.

 

[darkblu]

Re: ...

And I don't consider games to be the kind of software that rends itself to parallelism, there are too much inter-dependencies between game data structure to go MPP. Non-interactive graphics rendering, yes. Games, no.

pardon my curiosity, but how many games have you worked on?

 

[chaphack]

Nope. CELL won't change IBM, CELL won't change Sony, CELL will only affect SCEI. Come on, CELL powered TV sets and DVD players??? Who else is going to put SCEI's proprietary CELL in their product? Matsushita? JVC? Samsung? For what purpose? Will Sony want to carry a substantial cost burden that will serve little purpose??? Will consumers go gaga over $2000 TV sets that will supposely accelerate their PSX3 games by 10%??? Will IBM dump its X86 and POWER servers in favor of CELL servers??? Use some common sense here.

CELL, if successful, will make somewhat pretty game console. It won't change the world. Let's wake up from the dream and smell the coffee here.

Well said!

 

[maskrider]

Re: ...

Nope. CELL won't change IBM, CELL won't change Sony, CELL will only affect SCEI. Come on, CELL powered TV sets and DVD players??? Who else is going to put SCEI's proprietary CELL in their product? Matsushita? JVC? Samsung? For what purpose? Will Sony want to carry a substantial cost burden that will serve little purpose??? Will consumers go gaga over $2000 TV sets that will supposely accelerate their PSX3 games by 10%??? Will IBM dump its X86 and POWER servers in favor of CELL servers??? Use some common sense here.

CELL, if successful, will make somewhat pretty game console. It won't change the world. Let's wake up from the dream and smell the coffee here.

Yes, that's your opinion. And BTW, I don't remember anybody here said Cell will change the world. I only know that our biological cells are changing the world, at least the world around us.

But it will change at least Sony IMO, just need to understand that Cell is not a fabricated chip, it is a concept and an architecture.

Stay calm to discuss, or there is indeed no discussion and just arguing for the sake of arguing.

 

[Phil]

Nope. CELL won't change IBM, CELL won't change Sony, CELL will only affect SCEI. Come on, CELL powered TV sets and DVD players??? Who else is going to put SCEI's proprietary CELL in their product? Matsushita? JVC? Samsung? For what purpose? Will Sony want to carry a substantial cost burden that will serve little purpose??? Will consumers go gaga over $2000 TV sets that will supposely accelerate their PSX3 games by 10%??? Will IBM dump its X86 and POWER servers in favor of CELL servers??? Use some common sense here.

...and I'm not sure if anyone should bother arguing with someone as ignorant... :? Seriously, how can you argue CELL if you don't even see what the architecture/concept could do beyond just being a pretty CPU for the next PlayStation?

Since you're the one who wants 'us' to use common sense, how about using some of your own and ask yourself why Sony is willing to invest that much of money into a single concept/architecture? Then, think about Toshiba and IBM who are investing similar amounts...

 

[Vince]

Re: ...

Come on, CELL powered TV sets and DVD players??? Who else is going to put SCEI's proprietary CELL in their product? Matsushita? JVC? Samsung? For what purpose? Will Sony want to carry a substantial cost burden that will serve little purpose??? Will consumers go gaga over $2000 TV sets that will supposely accelerate their PSX3 games by 10%???

You really have no clue what you're talking about, it's sad. I'm going to recommend that you read up on Ando and his vision for Sony Group.

I won't wast my time going threw it all, but just dismiss your ignorant opinion of the future for TV's which you used as an example:

http://news.bbc.co.uk/1/hi/technology/2644589.stm[/url]]The first 50 years of colour television was just the infancy stage," he said during a keynote speech at the Consumer Electronics Show in Las Vegas. "The TV is going to be reborn as an always on and connected device."

The Japanese media giant is keen to develop the potential of the TV as a gadget to deliver its vast library of music, movies and TV shows to home viewers over the net.

It realises that few people are willing to watch a film on their computer.


TVs to be connected to the internet
The Tokyo company is working on new products designed to connect to the internet and to one another, allowing people to listen to music or watch films on their PC, TV or mobile phone.

"Sony's vision is a ubiquitous value network," said Mr Ando, "all devices connected so that you can enjoy content anytime, anywhere."

You are clueless... I'll leave it at that.

That's odd, since IBM came up with the term Cellular Computing in 1999 and Blue Gene project began right after Deep Blue.

I already explained the minimalist influence of Blue Gene on Cell. Thus unless you can dispute that (you can't) this is irrlevent. Thanks for wasting a post.

If Kutaragi's CELL was an independent idea why go to IBM and pay big bucks to have it developed??? After all, CELL is an IBM baby and SCEI is a paying customer, much like EE was a Toshiba baby and SCEI was the paying customer.

Because of the farraching scope of the project and the significant non-hardware related costs, in addition to IBM's superior R&D resources.

Cell's developers from IBM are mostly from their Server-Group and taken off Power4 development.

I already see the connection there. Big Blue doing research on something called "Cellular Computing", and Kutaragi san wanting a piece of it....

Um, I miss Chap.

Just compare them, Cell is much more coarsly grained in execution units and hierarchial - hardly Blue Gene like.

This is why I said that Kutaragi wanted to CELLtize EE VUs after seeing Blue Gene.

Ohh my... lets invoke Occam's Razor to shut you and your far-fetched and ultimatly fallicious theories up.

Dreamcast was completed by early 1998 and Kutaragi had its full spec by then. Of course he wanted to beat Sega numbers by a large margin, and this is why he ordered Toshiba to throw in VU1.

Bullshit, read the papers. You have ZERO proof of this... it's all in your mind. I posted 4-5 sources pinning the Emotion Engine as going public in early 1999. Development, if we utilize cell as a timeline/base, was begun in 1997.

You can't just produce a processor overnight. With the volume of publications in early 1997 describing the vector processors (eg. VU1 & VU0) it's impossible for your scenario to pann out. The papers I posted serve as fact... something your argument has no semblence of.